Home » Solutions » Private Data Cloud

GemStone in Private Data Cloud

GemStone in Private Data Cloud

GEMFIRE FOR CLOUD COMPUTING

“Many large enterprises have delayed broadening their grid deployments because of limitations in data management capabilities” – The 451 Group

Cloud/grid computing enables a model for the aggregation of all IT resources into an enterprise-wide virtual resource pool. Resources can then be dynamically provisioned across lines of business to accommodate fluctuating load and reliably fulfill demanding service-level agreements. However, “grid” often refers to a compute grid with no explicit reference to data. Consequently, most grid implementations are subject to significant data latency problems and cannot scale out to meet the demands the grid was originally designed to address. Modern day grid deployments require a new look at data architecture patterns. Traditional systems like relational database management systems cannot provide the necessary performance for distributed grid environments or enable data sharing. Without fast data access, most compute grids become fast-running processes with no data to process.

The GemFire Enterprise Data Fabric provides a best-in-class data grid implementation. GemFire virtualizes data from multiple sources and represents it in memory on physically distributed grid nodes. This delivers a single, extensible enterprise-wide distributed data grid that enables any process to reliably share, store, replicate, transform, route, and synchronize large volumes of data across the grid in real time. GemFire product lines provide native support for both Java and C++ applications and can be used in heterogeneous environments. Benefits of using the GemFire data grid include a multi-fold increase in throughput and performance without any extra investment in hardware.

GemFire data grids can augment compute grid environments by enabling data-aware task scheduling, sharing intermediate results between different steps in compute processes, and facilitating non-blocking task dependencies between compute tasks through a listener framework.